Can a Quantum Fourier Transform Level the Playing Field in MRI Systems?

Article Index

The second Electronics World article I would call to readers attention is from June, entitled: “3D Electronic Brain Atlas Model for the Detection of Neurological Disorders” from researchers at Selcuk University in Turkey. In their article they state:“for example, 3D MRI image analysis with the Freesurfer program can take 30 hours. Whereas the FSL program does volume analysis of only 15 subcortical regions it takes longer for analysis”.

Here the authors Muhammet Usame Ozic and Seral Ozsen utilize Open Access Data from Washington University. Helpfully they list some of the current software toolsets in common usage, not just from US institutions but also UK institutions, such as at Oxford and UCL (MATLAB based again). Albeit they don't inform us of the actual computational hardware processors.

As indicated in my previous Blog the question arises as to which of these specific Mathematical Toolsets shall survive undisrupted by Quantum Computing Augmentation of the 5G Forth Industrial revolution. Why so? In short: a technical and commercial way simply must be found to bring these vast computational overheads down from tens of hours to tens of seconds or tens of milliseconds even by harnessing them on an industrial scale through the Cloud. Consequently the limitation should not be computational throughput but the scan and dwell acquisition time of the individual MRI measurements.

Currently if you have a routine MRI scan on the NHS on the Monday in most cases it is the Thursday or Friday before you are presented with the results. This process is actually shortened somewhat in Cardiff, through globalisation, by sending the MRI measured data to Australia for overnight analysis (human).

These authors simple metrics determine the nature of the revolution required for real time Quantum Signal Processing and the entrenched inertia within the NHS, but alas also in “big dumb companies” as it has been said.

So it appears to me that it would be helpful, in the first instance, if our colleagues from Turkey could seek to establish what is the computational processing time saving to be achieved by performing their analysis on a Quantum Fourier Transform Algorithm instead. For in terms of algorithm complexity, only these researchers know what their actual MRI calculation pipeline is, with the divisions between the signal processing and the data processing. What I don't mean here, as one of my colleagues stated about a similar MRS algorithm: “it is of the order of 10^17 floating-point operations”. Wrong answer!

Once they have the true order of complexity identified, the next issue for them to address is sending the data to the IBM Q Cloud and back over current broadband pipelines and then the latency improvement factor 5G broadband can make. A separate but very relevant issue is whether they shall be permitted access to these IBM C Cloud based processors? And what to do next if they can't?

Just consider the installed global base of MRI scanners (GE Healthcare, Philips, Siemens etc), given the wake-up-call over the WannaCry Hack? Surely these legacy systems shall need to be continually upgraded to enjoy the significant computational enhancement with the wider adoption of Quantum Processing in the Cloud. Not just from IBM. But what particular new Fintech Business Model shall they need to adopted here? Simply put, physically, technically and commercially it shall not be possible for on site installation of a QC attached to every MRI Scanner. Instead all these installed legacy instruments shall need to be augmented to securely upload the raw unprocessed sampled data to the Cloud. Then securely download the processed images and assessments. This shall entail massive disruption not only to these instruments themselves but also to the way the real time MRI data traffic is managed and costed. Shall the above vendors provide this service? For big companies prefer to work with big companies and institutions, it is a cultural bonding issue for the employees themselves, but it is individuals and small companies and start-ups who make the real breakthroughs in innovation.

Here an Open Internet and Net Neutrality is of fundamental importance in leveling the playing field.
Just consider if the MRI scanner is in Turkey and the Quantum Fourier Processing or its Datod Substitute is performed in New York? Especially if the gate-keeping and routing Fintech Algorthms are accessed from within the UK.